Merge branch 'dev' into drop-python3.9

Author: joein

tests/congruence_tests/test_common.py

@@ -21,7 +21,6 @@
 # sparse vectors sizes
 sparse_text_vector_size = 100
 sparse_image_vector_size = 1_000
-sparse_code_vector_size = 10_000
 
 # number of vectors to generate
 NUM_VECTORS = 1000
@@ -44,7 +43,6 @@
 sparse_vectors_config = {
     "sparse-text": models.SparseVectorParams(),
     "sparse-image": models.SparseVectorParams(),
-    "sparse-code": models.SparseVectorParams(),
 }
 
 dense_vectors_sizes = {
@@ -56,7 +54,6 @@
 sparse_vectors_sizes = {
     "sparse-text": sparse_text_vector_size,
     "sparse-image": sparse_image_vector_size,
-    "sparse-code": sparse_code_vector_size,
 }
 
 multivectors_sizes = {

tests/congruence_tests/test_optional_vectors.py

@@ -131,7 +131,7 @@ def test_simple_opt_sparse_vectors_search():
                 COLLECTION_NAME,
                 vectors_to_retrieve,
                 with_payload=False,
-                with_vectors=["sparse-image", "sparse-code"],
+                with_vectors=["sparse-image", "sparse-text"],
             ),
             key=lambda x: x.id,
         ),
@@ -164,7 +164,7 @@ def test_simple_opt_sparse_vectors_search():
                 COLLECTION_NAME,
                 vectors_to_retrieve,
                 with_payload=False,
-                with_vectors=["sparse-image", "sparse-code"],
+                with_vectors=["sparse-image", "sparse-text"],
             ),
             key=lambda x: x.id,
         ),

tests/congruence_tests/test_query.py

@@ -20,7 +20,6 @@
     text_vector_size,
     sparse_text_vector_size,
     sparse_image_vector_size,
-    sparse_code_vector_size,
     generate_sparse_fixtures,
     sparse_vectors_config,
     generate_multivector_fixtures,
@@ -60,9 +59,6 @@ def __init__(self):
         self.sparse_vector_query_image = generate_random_sparse_vector(
             sparse_image_vector_size, density=0.2
         )
-        self.sparse_vector_query_code = generate_random_sparse_vector(
-            sparse_code_vector_size, density=0.1
-        )
 
         # multivector query vectors
         self.multivector_query_text = generate_random_multivector(text_vector_size, 3)
@@ -1661,6 +1657,31 @@ def test_query_group():
     init_client(local_client, secondary_collection_points, SECONDARY_COLLECTION_NAME)
     init_client(http_client, secondary_collection_points, SECONDARY_COLLECTION_NAME)
 
+    http_client.create_payload_index(
+        COLLECTION_NAME, field_name="id", field_schema=models.PayloadSchemaType.INTEGER
+    )
+    http_client.create_payload_index(
+        COLLECTION_NAME, field_name="rand_digit", field_schema=models.PayloadSchemaType.INTEGER
+    )
+    http_client.create_payload_index(
+        COLLECTION_NAME, field_name="two_words", field_schema=models.PayloadSchemaType.KEYWORD
+    )
+    http_client.create_payload_index(
+        COLLECTION_NAME,
+        field_name="city.name",
+        field_schema=models.PayloadSchemaType.KEYWORD,
+    )
+    http_client.create_payload_index(
+        COLLECTION_NAME,
+        field_name="maybe",
+        field_schema=models.PayloadSchemaType.KEYWORD,
+    )
+    http_client.create_payload_index(
+        COLLECTION_NAME,
+        field_name="maybe_null",
+        field_schema=models.PayloadSchemaType.KEYWORD,
+    )
+
     searcher.group_size = 5
     searcher.limit = 3
     for key in group_by_keys():
@@ -1687,6 +1708,7 @@ def test_query_group():
         )
 
     searcher.group_by = "city.name"
+
     for i in range(100):
         query_filter = one_random_filter_please()
         try:

tests/congruence_tests/test_retrieve.py

@@ -104,15 +104,15 @@ def test_sparse_retrieve() -> None:
     compare_client_results(
         local_client,
         remote_client,
-        lambda c: c.retrieve(COLLECTION_NAME, [id_], with_vectors=["sparse-image", "sparse-code"]),
+        lambda c: c.retrieve(COLLECTION_NAME, [id_], with_vectors=["sparse-image", "sparse-text"]),
     )
     compare_client_results(
         local_client,
         remote_client,
         lambda c: c.retrieve(
             COLLECTION_NAME,
             [id_],
-            with_vectors=["sparse-image", "sparse-code"],
+            with_vectors=["sparse-image", "sparse-text"],
             with_payload=False,
         ),
     )

tests/congruence_tests/test_sparse_recommend.py

@@ -47,7 +47,7 @@ def many_recommend(cls, client: QdrantBase) -> list[models.ScoredPoint]:
             query=models.RecommendQuery(recommend=models.RecommendInput(positive=[10, 19])),
             with_payload=True,
             limit=10,
-            using="sparse-image",
+            using="sparse-text",
         ).points
 
     @classmethod
@@ -59,7 +59,7 @@ def simple_recommend_negative(cls, client: QdrantBase) -> list[models.ScoredPoin
             ),
             with_payload=True,
             limit=10,
-            using="sparse-image",
+            using="sparse-text",
         ).points
 
     @classmethod
@@ -106,7 +106,7 @@ def best_score_recommend(cls, client: QdrantBase) -> list[models.ScoredPoint]:
         ).points
 
     @classmethod
-    def best_score_recommend_euclid(cls, client: QdrantBase) -> list[models.ScoredPoint]:
+    def best_score_recommend_pos_neg(cls, client: QdrantBase) -> list[models.ScoredPoint]:
         return client.query_points(
             collection_name=COLLECTION_NAME,
             query=models.RecommendQuery(
@@ -118,7 +118,7 @@ def best_score_recommend_euclid(cls, client: QdrantBase) -> list[models.ScoredPo
             ),
             with_payload=True,
             limit=10,
-            using="sparse-code",
+            using="sparse-image",
         ).points
 
     @classmethod
@@ -135,22 +135,6 @@ def only_negatives_best_score_recommend(cls, client: QdrantBase) -> list[models.
             using="sparse-image",
         ).points
 
-    @classmethod
-    def only_negatives_best_score_recommend_euclid(
-        cls, client: QdrantBase
-    ) -> list[models.ScoredPoint]:
-        return client.query_points(
-            collection_name=COLLECTION_NAME,
-            query=models.RecommendQuery(
-                recommend=models.RecommendInput(
-                    positive=None, negative=[10, 12], strategy=models.RecommendStrategy.BEST_SCORE
-                )
-            ),
-            with_payload=True,
-            limit=10,
-            using="sparse-code",
-        ).points
-
     @classmethod
     def sum_scores_recommend(cls, client: QdrantBase) -> list[models.ScoredPoint]:
         return client.query_points(
@@ -166,7 +150,7 @@ def sum_scores_recommend(cls, client: QdrantBase) -> list[models.ScoredPoint]:
         ).points
 
     @classmethod
-    def sum_scores_recommend_euclid(cls, client: QdrantBase) -> list[models.ScoredPoint]:
+    def sum_scores_recommend_pos_neg(cls, client: QdrantBase) -> list[models.ScoredPoint]:
         return client.query_points(
             collection_name=COLLECTION_NAME,
             query=models.RecommendQuery(
@@ -178,7 +162,7 @@ def sum_scores_recommend_euclid(cls, client: QdrantBase) -> list[models.ScoredPo
             ),
             with_payload=True,
             limit=10,
-            using="sparse-code",
+            using="sparse-image",
         ).points
 
     @classmethod
@@ -195,22 +179,6 @@ def only_negatives_sum_scores_recommend(cls, client: QdrantBase) -> list[models.
             using="sparse-image",
         ).points
 
-    @classmethod
-    def only_negatives_sum_scores_recommend_euclid(
-        cls, client: QdrantBase
-    ) -> list[models.ScoredPoint]:
-        return client.query_points(
-            collection_name=COLLECTION_NAME,
-            query=models.RecommendQuery(
-                recommend=models.RecommendInput(
-                    positive=None, negative=[10, 12], strategy=models.RecommendStrategy.SUM_SCORES
-                )
-            ),
-            with_payload=True,
-            limit=10,
-            using="sparse-code",
-        ).points
-
     @classmethod
     def avg_vector_recommend(cls, client: QdrantBase) -> list[models.ScoredPoint]:
         return client.query_points(
@@ -286,7 +254,6 @@ def recommend_batch(client: QdrantBase) -> list[models.QueryResponse]:
 
 def test_simple_recommend() -> None:
     fixture_points = generate_sparse_fixtures()
-
     secondary_collection_points = generate_sparse_fixtures(100)
 
     searcher = TestSimpleRecommendation()
@@ -320,35 +287,26 @@ def test_simple_recommend() -> None:
         vectors_config={},
         sparse_vectors_config=sparse_vectors_config,
     )
-
     compare_client_results(local_client, remote_client, searcher.simple_recommend_image)
     compare_client_results(local_client, remote_client, searcher.many_recommend)
     compare_client_results(local_client, remote_client, searcher.simple_recommend_negative)
     compare_client_results(local_client, remote_client, searcher.recommend_from_another_collection)
     compare_client_results(local_client, remote_client, searcher.best_score_recommend)
-    compare_client_results(local_client, remote_client, searcher.best_score_recommend_euclid)
+    compare_client_results(local_client, remote_client, searcher.best_score_recommend_pos_neg)
     compare_client_results(
         local_client, remote_client, searcher.only_negatives_best_score_recommend
     )
-    compare_client_results(
-        local_client, remote_client, searcher.only_negatives_best_score_recommend_euclid
-    )
     compare_client_results(local_client, remote_client, searcher.sum_scores_recommend)
-    compare_client_results(local_client, remote_client, searcher.sum_scores_recommend_euclid)
+    compare_client_results(local_client, remote_client, searcher.sum_scores_recommend_pos_neg)
     compare_client_results(
         local_client, remote_client, searcher.only_negatives_sum_scores_recommend
     )
-    compare_client_results(
-        local_client, remote_client, searcher.only_negatives_sum_scores_recommend_euclid
-    )
-
     compare_client_results(local_client, remote_client, searcher.avg_vector_recommend)
     compare_client_results(local_client, remote_client, searcher.recommend_from_raw_vectors)
     compare_client_results(
         local_client, remote_client, searcher.recommend_from_raw_vectors_and_ids
     )
     compare_client_results(local_client, remote_client, searcher.recommend_batch)
-
     for _ in range(10):
         query_filter = one_random_filter_please()
         try:

tests/congruence_tests/test_sparse_search.py

@@ -11,7 +11,6 @@
     init_client,
     init_local,
     init_remote,
-    sparse_code_vector_size,
     sparse_image_vector_size,
     sparse_text_vector_size,
     sparse_vectors_config,
@@ -26,7 +25,6 @@ class TestSimpleSparseSearcher:
     def __init__(self):
         self.query_text = generate_random_sparse_vector(sparse_text_vector_size, density=0.3)
         self.query_image = generate_random_sparse_vector(sparse_image_vector_size, density=0.2)
-        self.query_code = generate_random_sparse_vector(sparse_code_vector_size, density=0.1)
 
     def simple_search_text(self, client: QdrantBase) -> list[models.ScoredPoint]:
         return client.query_points(
@@ -48,16 +46,6 @@ def simple_search_image(self, client: QdrantBase) -> list[models.ScoredPoint]:
             limit=10,
         ).points
 
-    def simple_search_code(self, client: QdrantBase) -> list[models.ScoredPoint]:
-        return client.query_points(
-            collection_name=COLLECTION_NAME,
-            using="sparse-code",
-            query=self.query_code,
-            with_payload=True,
-            with_vectors=True,
-            limit=10,
-        ).points
-
     def simple_search_text_offset(self, client: QdrantBase) -> list[models.ScoredPoint]:
         return client.query_points(
             collection_name=COLLECTION_NAME,
@@ -122,7 +110,7 @@ def simple_search_image_select_vector(self, client: QdrantBase) -> list[models.S
             using="sparse-image",
             query=self.query_image,
             with_payload=False,
-            with_vectors=["sparse-image", "sparse-code"],
+            with_vectors=["sparse-image", "sparse-text"],
             limit=10,
         ).points
 
@@ -186,7 +174,6 @@ def test_simple_search():
 
     compare_client_results(local_client, remote_client, searcher.simple_search_text)
     compare_client_results(local_client, remote_client, searcher.simple_search_image)
-    compare_client_results(local_client, remote_client, searcher.simple_search_code)
     compare_client_results(local_client, remote_client, searcher.simple_search_text_offset)
     compare_client_results(local_client, remote_client, searcher.search_score_threshold)
     compare_client_results(local_client, remote_client, searcher.simple_search_text_select_payload)
@@ -235,7 +222,6 @@ def test_simple_opt_vectors_search():
 
     compare_client_results(local_client, remote_client, searcher.simple_search_text)
     compare_client_results(local_client, remote_client, searcher.simple_search_image)
-    compare_client_results(local_client, remote_client, searcher.simple_search_code)
     compare_client_results(local_client, remote_client, searcher.simple_search_text_offset)
     compare_client_results(local_client, remote_client, searcher.search_score_threshold)
     compare_client_results(local_client, remote_client, searcher.simple_search_text_select_payload)